Implementation methods of infection prevention measures in orthopedics and traumatology – a systematic review

Prevention of hospital-acquired infections is one of the major concerns of patients and physicians alike. Perioperative infections in the clinical field of orthopedics and traumatology pose a huge obstacle as the consequences in case of osteitis and implant failure can be devastating [1‐3]. Implant-related infections usually require long-term antibiotic therapy with possible development of resistances and possibly multiple surgical revisions [4, 5]. In addition to clinical outcomes, the negative socioeconomic implications are immense [5‐7]. There has been an increasing number of publications concerning successful single and bundled infection prevention measures in recent years. Those summarizing, the World Health Organization published a guideline on preventive measures of surgical site infections in 2016 [8]. In addition, reviews on specific IP measures in the clinical field of orthopedics and traumatology were published which surpassed the WHO recommendations [9, 10]. Although in many cases, we do theoretically know which IP measures are associated with the best outcome, the measures themselves are often carried out insufficiently in practice [11, 12]. The question remains: Which methods/strategies to implement those proven IP measures into our clinical practice work best? The relatively young field of implementation science aims to find scientifically proven ways to transmit desired operations into practice [13].

As part of the HYGArzt-Project (“Proof Of Effectivity And Efficiency Of Implementation Of Infection Prevention (IP) Measures By The Physician Responsible For Infection Prevention Matters In Traumatology/Orthopedics” Grant Number ZMVI1‐2516FSB111), the objective of this study was to systematically review and identify effective implementation methods of IP measures in orthopedics and traumatology.

Evidence for implementation methods of IP measures in orthopedics and trauma surgery is scarce. Nevertheless, we were able to identify some recurring implementation methods that seem to be valuable when planning IP improvement interventions. Interpersonal communication and interdisciplinary work seem to play a role in successful implementation. It, thus, seems recommendable to adhere to following methods as a standardized process when planning and executing an intervention with the aim of implementing IP measures in orthopedics and traumatology:

These implementation methods address behavioral as well as cognitive levels; the multilevel bundled approach proves valuable for IP measures as well as implementation methods. The interdisciplinary approach has to be stressed as the danger of therapy-associated infection is an interdisciplinary, multifaceted problem. Our findings suggest the involvement of surgeons as primary stakeholders and champions of IP and implementation. Nursing professions should be given a forum as they carry out a big portion of the workload. On the other hand, the executive suite, management professions, economic and clinical leaders should be involved to stabilize planning, made easier by interdisciplinary agreeing on guidelines and best practice to estimate costs.

It is noteworthy that every study on the implementation of IP measures comprised by more than one implementation method reported clinical success. This means that, apart from economic considerations, as long as one does intervene—there will likely be success [36].

The data extracted in the review cannot provide a quantification of the importance of each single implementation method, but some of the included studies reported what the authors stressed as important factors for success. Multidisciplinary team approach and interdisciplinary consensus were mentioned in 6 out of 11 successful studies [21, 23, 24, 27‐29]. Other aspects mentioned as impactful were personal training [19, 29], continuous training and feedback [24, 29], monthly analysis [21], active involvement of every member and physician leadership [21, 24] and patience [24]. Implementation methods reaching the cognitive level only (education by posters in this case) were not followed by improvement, which fits established implementation research [30]. The impression gained from the included studies is that there should be no separate entities in terms of IP measure and implementation method but rather a synthesis of what and how: The how-aspect of infection prevention seems to pose the greater obstacle for clinicians at times, accordingly clinical research on infection prevention should automatically also be researched on clinical implementation. The results go along well with current implementation science´s findings and theoretical models, many of which were published later than the studies included in this review [31, 32]. Beyond our own findings in orthopedics and trauma surgery, implementation science has gathered knowledge about change management in the health care sector. Five major domains were identified as being crucial for implementation: outer setting, for example the resources of a healthcare system; inner setting, for example, the structure of the ward where the implementation takes place; characteristics of the individuals involved; intervention characteristics; and the process of implementation [31]. According to our results, implementation measures used in trauma surgery so far focus on subdomains of the inner setting (teambuilding aspects) and the process of implementation itself (how to inform staff, measurement and feedback of results). Since implementation barriers are often based on problems with team communication, and departmental culture is a key factor for successful implementation [33], giving special attention to team building measures seems like a reasonable approach. Members must be aware of being a part of the implementation process [33]. Nevertheless, in future studies, additional domains or subdomains may be taken into consideration to enhance the overall effect and implementation success. Leadership and the support of management and middle management play an important role [31, 33, 34]. The legitimacy of the source may also influence implementation [31]. It seems to play a role if an intervention is initiated by a widely accepted entity. Following the Theory of planned behaviour by Ajzen, human behaviour depends on attitude, subjective norm and perceived behavioural control [35]. Thus, besides convincing the team members and conveying of content, etc., it is crucial to make the desirable conduct as easy and comfortable as possible for the staff. This may be another factor to increase compliance.

There are, however, limitations to the recommendations both for formal reasons and with regard to contents. Primarily, there are concerns about the quality of the included studies.

The MINORS score average is rather low with a probably moderate-to-high risk of bias. After quality assessment via MINORS scores and Risk of Bias evaluation, we decided to keep all studies in the review despite the disparate range of 4–13 out of 16 points, as there is no clear cut-off suggestion on MINORS scores found in the literature and we decided to rather discuss the findings in a cautious way. Since the results of the studies reporting success are rather homogeneous, we did not see an additional benefit for the interpretation by weighing the individual Risk of bias of the included studies. Interventions designed to achieve behavioural change or change in overall culture of a workplace surrounding are inherently biased as in that every possible parameter is manipulated towards achieving the desired outcome, only the outcome itself should in the end be measured neutrally. The studies not showing successful intervention showed a clear enough difference in terms of implementation methods used not to conclude a heterogenity of results. Funding of the included studies was not consistently reported with only 5 out of 13 studies claiming no external funding or public funding. The inconsistency reporting of data made data pooling unfeasible. Reported improvements, therefore, have to be processed cautiously. In regard to contents, it has to be noted that rather frequently there was no in-depth description of implementation methods. There is, for example, no universal definition of interactive training or personal training or teaching [32, 36]. This terminology was inconsistent throughout the studies, which has been criticized in implementation research before; therefore, we had to sub-summarize these methods as one entity, though it might well play an important role how exactly the interpersonal work was designed [37]. Occasionally, the clear separation of an IP measure and an implementation method was difficult (which leads to the above stated conclusion that generally they should not be considered separate entities). Additionally, it was not possible to determine which of the implementation methods included in the successful studies had the biggest impact, which is a known problem, but it was possible to approximate by analyzing the frequency of use [36, 38]. Addressing the problems encountered (and possible solutions) while implementing IP measures could also be underreported and possibly biased. One author stated that they underestimated the “complexity of hospital structures and surgical outcome measures” as well as the “variability of perceived role of disciplines” [28]. Another author faced economic problems as the implementation cost per patient seemed to exceed the cost of one SSI [27]. One study reported limited resources for the pharmacy department as advisors to the surgical department as a hurdle [19]. Two authors reported a feedback from implementers stating a “lack of physician buy-in and staff resistance”, economic aspects (cost of antibiotics/antiseptic agents), logistics of preoperative IP measures (antiseptic washings) and “a perceived lack of evidence” were the main problem fields [17, 18]. One hospital representative selected for a QI campaign stated that “SSIs are not a problem” [17, 18]. One study discussed that adherence and improvement thereof would be measurable, but sample sizes to be underpowered to demonstrate an effect on SSI% [23].

There might be important insight on the topic of implementation of IP measures from other clinical fields which was not considered in this study due to the limitation to orthopedics and traumatology [39‐50].

Although evidence was scarce and quality-inconsistent, we found that adhering to a set of implementation methods focusing on interdisciplinary and interactive/interpersonal work might be an advisable strategy when planning IP improvement interventions in orthopedics and traumatology.